Pith. sign in

REVIEW 2 cited by

Decentralized Deep Learning with Arbitrary Communication Compression

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1907.09356 v3 pith:5EWHP2DI submitted 2019-07-22 cs.LG cs.DCcs.DSmath.OCstat.ML

classification cs.LGcs.DCcs.DSmath.OCstat.ML
keywords compressionlearningdecentralizeddeepnumbertrainingalgorithmarbitrary
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

Decentralized training of deep learning models is a key element for enabling data privacy and on-device learning over networks, as well as for efficient scaling to large compute clusters. As current approaches suffer from limited bandwidth of the network, we propose the use of communication compression in the decentralized training context. We show that Choco-SGD $-$ recently introduced and analyzed for strongly-convex objectives only $-$ converges under arbitrary high compression ratio on general non-convex functions at the rate $O\bigl(1/\sqrt{nT}\bigr)$ where $T$ denotes the number of iterations and $n$ the number of workers. The algorithm achieves linear speedup in the number of workers and supports higher compression than previous state-of-the art methods. We demonstrate the practical performance of the algorithm in two key scenarios: the training of deep learning models (i) over distributed user devices, connected by a social network and (ii) in a datacenter (outperforming all-reduce time-wise).

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Decentralized Optimization with Amplified Privacy via Efficient Communication

    eess.SY 2025-06 reject novelty 6.0 of 10

    Random activation and Top-k sparsification are claimed to amplify differential privacy in decentralized non-convex optimization, reducing required noise by a factor of the sparsification ratio times the square of the ...

  2. A Communication-Efficient Distributed Optimization Algorithm for Problems with Coupling Constraints

    math.OC 2025-12 reject novelty 5.0 of 10

    A compressed dual-splitting algorithm with dynamic scaling is claimed to converge linearly under unbiased and biased quantizers, but the main proof identity is false.

Pith tools